Transformer oil particle online monitoring device based on spectrum and conductivity detection

Abstract

The invention provides an on-line monitoring device for transformer oil particles based on spectrum and conductivity detection, which realizes efficient, accurate and continuous monitoring of particle impurities in transformer oil. The system comprises a sampling module, a sensor module, an analysis processing unit and a communication module. The sampling module comprises a micro pump for realizing continuous oil sample collection and an adjustable throttle valve for ensuring the stability of an oil sample; the sensor module comprises a spectrum sensor for detecting optical characteristics in an oil sample and a conductivity sensor for monitoring conductivity change of the oil sample and detecting electrical characteristics of different impurities in oil; the microprocessor is used for processing sensor data and judging impurity types and concentrations; and the communication module adopts a wireless transmission technology to send the impurity type and concentration obtained by the analysis processing unit to a monitoring center in real time. The device is simple and convenient to operate, accurate and reliable in detection result and suitable for on-line monitoring and maintenance of transformer oil.

Description

Technical Field

[0001] The invention relates to the technical field of power transformer detection, and in particular to an online transformer oil particle monitoring device based on spectrum and conductivity detection. Background Art

[0002] During the operation of power transformers, the insulating oil will be contaminated by various impurities. These impurities include tiny particles, moisture, gas and other pollutants, which will have a negative impact on the performance of the insulating oil. In particular, impurity particles, these tiny particles will significantly affect the insulation performance of the transformer, reduce its withstand voltage level, increase the risk of partial discharge, and thus threaten the safe and stable operation of the transformer. People have long recognized the importance of keeping insulating oil clean, but there are many problems with existing detection methods.

[0003] Traditional detection methods usually rely on regular sampling and laboratory analysis, which is not only complex and inefficient, but also cannot reflect the status of transformer oil in real time. In actual applications, the operating environment of transformers is complex and changeable, and the impurity content in the oil may change significantly in a short period of time. If these changes cannot be detected and handled in time, it may cause transformer failure and even cause safety accidents in the power system.

[0004] The existing detection methods of impurity particles in transformer oil mainly rely on offline detection technology, which usually requires multiple steps such as sampling, transportation, and laboratory analysis. The process is cumbersome and time-consuming. Because it is an offline detection, it is difficult to reflect the changes of impurity particles in transformer oil in a timely manner, and it cannot meet the real-time monitoring requirements of the transformer operation status. In addition, the existing detection equipment is usually large in size, expensive, and has high maintenance costs, which is not suitable for field applications. And it requires professional personnel to operate, which increases labor costs and operational risks. Therefore, it is particularly important to develop a device that can detect impurity particles in transformer oil in real time online. Such a device should have high sensitivity and high accuracy, and can continuously monitor the state of insulating oil without stopping the machine, and timely discover and warn potential fault hazards. This can not only improve the operating reliability of the transformer, but also extend its service life and reduce the cost of maintenance and overhaul.

[0005] Through real-time online monitoring devices, the safety and efficiency of transformer operation will be significantly improved. The application of this technology will have a profound impact on the stable operation of the power system and is an important direction for the development of modern power equipment monitoring technology. Summary of the invention

[0006] In order to solve the deficiencies in the prior art, the present invention aims to provide an online monitoring device for particles in transformer oil, which can realize real-time detection of different impurity particles in oil and has the advantages of low maintenance cost and simple operation.

[0007] To achieve the above object, the present invention adopts the following technical solution: an online monitoring device for transformer oil particles based on spectrum and conductivity detection, the device comprising: a sampling module, a sensor module, an analysis and processing unit and a communication module.

[0008] The sampling module realizes continuous oil sample collection through a micro pump and an adjustable flow valve, and can provide oil samples stably and continuously; the sensor module includes a spectral sensor and a conductivity sensor, which are used to detect the optical and electrical properties in the oil sample, and improves the accuracy and comprehensiveness of the detection through multi-angle and multi-indicator detection methods; the spectral sensor is used to detect the optical properties in the oil sample, and the conductivity sensor is used to monitor the conductivity changes of the oil sample and detect the electrical properties of different impurities in the oil. The analysis and processing unit integrates a microprocessor, which can quickly process and analyze sensor data and accurately determine the type and concentration of impurities in the oil sample; the communication module uses wireless transmission technology to send the processed data to the monitoring center in real time to ensure the real-time and reliability of data transmission.

[0009] Furthermore, the integrated microprocessor includes an impurity type and concentration discrimination method, which includes a discrimination calculation model. The discrimination calculation model includes a classification module and a concentration fitting module, wherein the classification module is used to determine the impurity type in the oil sample, and the concentration fitting module is used to calculate the impurity concentration corresponding to the impurity under the determined impurity type, thereby outputting the impurity type and concentration.

[0010] The discrimination calculation model includes an offline training phase and an online detection phase.

[0011] Offline training stage: Collect a number of offline data, including spectral data obtained by detecting different impurities in transformer oil through a spectral sensor and conductivity data obtained by detecting transformer oil under different impurity concentration conditions through a conductivity sensor, and label the spectral data and conductivity data of each impurity;

[0012] The spectral data were processed using the multivariate linear regression algorithm and the principal component analysis algorithm to calculate the concentration of each different impurity, recorded as impurity concentration one; the conductivity data were processed using the gradient descent algorithm to calculate the concentration of each different impurity, recorded as impurity concentration two. Subsequently, the impurity concentration one and the impurity concentration two corresponding to each impurity were fused to obtain the final concentration of each impurity; the several final concentrations of each impurity were used as training set one, and the types of impurities in training set one were used as labels for type training samples.

[0013] A number of conductivity data of each impurity are used as training set 2, and the actual concentration of each impurity in training set 2 is used as the label of the corresponding concentration training sample;

[0014] The classification module is trained through training set one, and when the loss no longer decreases, a trained classification module is obtained. The concentration fitting module is trained through training set two, and when the loss no longer decreases, a trained concentration fitting module is obtained. After the classification module and concentration fitting training are completed, a trained discrimination calculation model is obtained.

[0015] Online detection stage: The data information collected by the spectral sensor and the conductivity sensor is input into the trained discrimination calculation model to output the types and concentrations of different particles.

[0016] Furthermore, in the offline training stage, the classification module uses the support vector machine (SVM) algorithm to construct a hyperplane to separate different categories of impurity data and minimize the hinge loss function. The concentration fitting module uses a regression analysis algorithm to optimize the model parameters through the least squares method or gradient descent method to ensure that the relationship between conductivity data and impurity concentration is accurately fitted.

[0017] The category training samples are input into the classification module, and the support vector machine algorithm SVM separates the data of different categories by constructing a hyperplane. The hyperplane is expressed as:

[0018] w·x+b=0 (1)

[0019] Among them, w is the normal vector, b is the bias term, and x is the input data.

[0020] SVM uses the hinge loss function, which is defined as:

[0021]

[0022] in is the i-th category training sample in training set 1 The predicted value of Represents the category label of the i-th category training sample.

[0023] The goal of classification module training is to minimize the following loss function:

[0024]

[0025] Among them, C is the regularization parameter, which is used to control the degree of fit between the complexity of the classification module and the training set, and n represents the number of category training samples in the training set.

[0026] The concentration fitting module uses a regression analysis algorithm to calculate the concentration of each impurity:

[0027]

[0028] Where Y is the concentration of each impurity, represents the j-th concentration training sample in training set 2, β0,β1,…,β m is the regression coefficient and ε is the error term.

[0029] The goal of concentration fitting module training is to minimize the loss function J(θ):

[0030]

[0031] Where m is the total number of concentration training samples in training set 2, is the actual impurity concentration of the j-th concentration training sample, is the predicted value of the concentration fitting module.

[0032] During the training process, the gradient descent algorithm is used to optimize the parameters of the classification module and the concentration fitting module, and to update the weight vector of the SVM and the coefficient of the regression analysis algorithm. The formula is as follows:

[0033]

[0034] Among them, θ is the model parameter, which is the weight vector in SVM and the regression coefficient in regression analysis algorithm; α is the learning rate, is the gradient of the loss function with respect to the parameters.

[0035] Furthermore, in the offline training stage, the method of using the multivariate linear regression algorithm and the principal component analysis algorithm to process the spectral data to calculate the concentration of different impurities is:

[0036] D concentration2 =X·β+ε1 (7)

[0037] Among them, D concentration1 is the impurity concentration, X is the spectral data obtained by the spectral sensor, β is the regression coefficient, and ε1 is the error term.

[0038] Furthermore, in the offline training stage, the conductivity data is processed using a gradient descent algorithm to calculate the concentration of different impurities. The method is:

[0039] The concentration fitting model uses a linear regression algorithm to fit the relationship between conductivity data and impurity concentration:

[0040] D concentration2 =β0+β1·x1+β2·x2+…+β n ·x n (8)

[0041] Among them, D concentration2 is the impurity concentration 2, x1, x2,…, x n is the conductivity data obtained by the conductivity sensor, β0, β1,…, βn is the regression coefficient, and the regression coefficient is obtained by least squares method or gradient descent algorithm in the offline training stage.

[0042] Furthermore, in the offline training stage, the concentrations of different impurities calculated by the spectral sensor data are calculated as concentration1 The concentration calculated from the conductivity sensor data is D concentration2 Perform fusion to obtain the final concentration of each impurity. The fusion formula is as follows:

[0043] D combined =α·D concentration1 +(1-α)·D concentration2 (9)

[0044] Among them, D combined represents the final concentration of each impurity, and α is the fusion weight.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention can detect the impurity characteristics in transformer oil from multiple angles and multiple indicators through the combination of sampling module, spectral sensor and conductivity sensor. Specifically, the spectral sensor can detect the optical characteristics of the oil sample and identify different types of impurities; the conductivity sensor can detect the electrical characteristics of the oil sample and reflect the content of conductive impurities in the oil sample. Through this multi-dimensional detection method, the accuracy and comprehensiveness of the detection are greatly improved, and more detailed oil sample analysis data can be provided.

[0047] 2. The communication module of the present invention adopts advanced wireless transmission technology and supports Wi-Fi and LoRa communication modes. Wireless transmission technology can ensure that the data is sent to the monitoring center in real time after processing, so that the monitoring center can receive and process the oil sample test data in real time, ensuring the real-time and reliability of data transmission. This real-time data transmission capability enables the monitoring center to grasp the operating status of the transformer in a timely manner, effectively prevent and promptly handle potential faults.

[0048] 3. The analysis and processing unit of the present invention integrates a high-performance microprocessor (model: MCU-500) and an impurity type and concentration discrimination method. The microprocessor has efficient data processing capabilities and can quickly process a large amount of sensor data. The impurity type and concentration discrimination method can accurately analyze the sensor data, determine the type and concentration of impurities in the oil sample, and thus generate a reliable test report. This efficient data processing capability not only improves the detection speed, but also ensures the accuracy and reliability of the test results.

[0049] 4. The sampling module of the present invention adopts a micro pump (MP-100) and an adjustable flow valve (FV-200) to achieve stable and continuous oil sample collection. The micro pump has a flow range of 0.1-1L / min, which can ensure stable oil sampling under different working conditions. The adjustable flow valve can accurately control the flow rate through the manual adjustment knob to ensure the stability and continuity of the oil sample, providing high-quality oil samples for sensor detection.

[0050] 5. Each module of the present invention is compact in design and easy to install at the sampling port of the transformer oil tank. The installation and debugging process is simple and convenient, and can ensure a stable connection without leakage. The operating status of each module is easy to check and maintain, and the user can recalibrate the sensor as needed to maintain the detection accuracy. The working condition of the micro pump and flow valve can be regularly checked and maintained, and replaced in time to ensure the continuity and stability of oil sample collection.

[0051] 6. The present invention can detect not only the optical properties of the oil sample, but also the electrical properties of the oil sample. Through the comprehensive processing of the analysis and processing unit, a detailed detection report can be generated to provide comprehensive monitoring and analysis for the operation status of the transformer. Different from the traditional single detection method, the present invention provides a more comprehensive and multifunctional oil sample detection solution, which improves the overall level of transformer oil detection.

[0052] 7. Existing detection methods have the problems of being unable to detect in real time online, complicated and time-consuming detection steps, high maintenance costs, and complex operations. These deficiencies make it difficult to meet the demand for efficient and accurate monitoring of impurity particles in transformer oil. The detection device of the present invention can not only detect impurity particles in oil in real time, but also identify different types of particles. It is easy to operate and has low maintenance costs, and has broad application prospects and market value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0054] Figure 1 The figure is a block diagram of the transformer oil particle online monitoring device of the present invention.

[0055] Figure 2 It is the principle of particle type and concentration detection.

[0056] Figure 3 This is a schematic diagram of the installation location of the detection device.

[0057] Among them, 1-transformer oil tank; 2-transformer insulating oil; 3-impurity particles; 4-transformer coil; 5-sampling module. DETAILED DESCRIPTION

[0058] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0059] It should be noted that the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. "If below", the meaning of "one or more", unless otherwise clearly and specifically defined.

[0060] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0061] An online monitoring device for transformer oil particles based on spectroscopy and conductivity detection includes the following components:

[0062] Sampling module: including a micro pump and an adjustable flow valve, wherein the micro pump is used to realize continuous oil sample collection, and the adjustable flow valve is used to ensure the stability of the oil sample;

[0063] Sensor module: including a spectral sensor and a conductivity sensor, wherein the spectral sensor is used to detect the optical properties in the oil sample, and the conductivity sensor is used to monitor the conductivity change of the oil sample and detect the electrical properties of different impurities in the oil;

[0064] Analysis and processing unit: integrated microprocessor method, the microprocessor is used to process sensor data and determine the type and concentration of impurities.

[0065] Communication module: wireless transmission technology is used to send the impurity type and concentration obtained by the analysis and processing unit to the monitoring center in real time.

[0066] In specific implementation, the sampling module includes a micro pump of model MP-100 and an adjustable flow valve of model FV-200. The micro pump has a flow range of 0.1-1L / min and can continuously and stably extract oil samples. The coordinated use of the micro pump MP-100 and the adjustable flow valve FV-200 ensures the stability and continuity of oil sample collection, providing high-quality samples for subsequent sensor detection.

[0067] In specific implementation, the sensor module includes a spectral sensor of model SP-300 and a conductivity sensor of model EC-100. The wavelength range of the spectral sensor is 200-800nm. The spectral sensor of SP-300 is used to detect the spectral characteristics in the oil sample. The spectral sensor SP-300 can identify different types of impurities by analyzing the spectral response of the oil sample and provide detailed optical detection data. The conductivity sensor EC-100 is used to detect the conductivity change of the oil sample. By monitoring the conductivity of the oil sample, the conductivity sensor EC-100 can reflect the content of conductive impurities in the oil sample and provide supplementary electrical detection data. The measurement range of the conductivity sensor is 0-200μS / cm, by monitoring the conductivity change of the oil sample.

[0068] In specific implementation, the analysis and processing unit integrates a microprocessor MCU-500 and an impurity type and concentration discrimination method to process sensor data and determine the type and concentration of impurities in the oil sample. The efficient computing power and dedicated algorithm of the microprocessor MCU-500 can accurately analyze the sensor data, determine the type and concentration of impurities in the oil sample, and provide reliable test results. The microprocessor has high-performance computing power, can quickly process large amounts of data, and accurately identify impurities in the oil through built-in algorithms to generate reliable test reports.

[0069] In specific implementation, the communication module model is WT-400, which supports Wi-Fi and LoRa communication modes, ensuring real-time transmission and reliable reception of data, so that the monitoring center can timely grasp the operating status of the transformer. WT-400 sends the impurity type and concentration obtained by the analysis and processing unit to the monitoring center in real time. The wireless transmission module WT-400 supports multiple communication modes, ensuring real-time transmission and reliable reception of data, so that the monitoring center can timely grasp the operating status of the transformer.

[0070] Specifically, the integrated microprocessor includes an impurity type and concentration discrimination method, which includes a discrimination calculation model. The discrimination calculation model includes a classification module and a concentration fitting module, wherein the classification module is used to determine the impurity type in the oil sample, and the concentration fitting module is used to calculate the impurity concentration corresponding to the impurity under the determined impurity type, thereby outputting the impurity type and concentration.

[0071] The discrimination calculation model includes an offline training phase and an online detection phase;

[0072] Offline training stage: Collect a number of offline data, including spectral data obtained by detecting different impurities in transformer oil through a spectral sensor and conductivity data obtained by detecting transformer oil under different impurity concentration conditions through a conductivity sensor, and label the spectral data and conductivity data of each impurity;

[0073] The spectral data were processed using the multivariate linear regression algorithm and the principal component analysis algorithm to calculate the concentration of each different impurity, recorded as impurity concentration one; the conductivity data were processed using the gradient descent algorithm to calculate the concentration of each different impurity, recorded as impurity concentration two. Subsequently, the impurity concentration one and the impurity concentration two corresponding to each impurity were fused to obtain the final concentration of each impurity; the several final concentrations of each impurity were used as training set one, and the types of impurities in training set one were used as labels for type training samples.

[0074] A number of conductivity data of each impurity are used as training set 2, and the actual concentration of each impurity in training set 2 is used as the label of the corresponding concentration training sample;

[0075] The classification module is trained through training set one, and when the loss no longer decreases, a trained classification module is obtained. The concentration fitting module is trained through training set two, and when the loss no longer decreases, a trained concentration fitting module is obtained. After the classification module and concentration fitting training are completed, a trained discrimination calculation model is obtained.

[0076] Online detection stage: The data information collected by the spectral sensor and the conductivity sensor is input into the trained discrimination calculation model to output the types and concentrations of different particles.

[0077] Specifically, in the model training phase, a large amount of offline data needs to be collected first. These training data include spectral data and conductivity data detected by spectral sensors and conductivity sensors, covering different concentrations of common impurities in transformer oil, and corresponding to known impurity types and their concentration labels.

[0078] In the offline training stage, the classification module uses the support vector machine algorithm SVM to construct a hyperplane to separate different types of impurity data and minimize the hinge loss function. The concentration fitting module uses a regression analysis algorithm to optimize the model parameters through the least squares method or gradient descent method to ensure that the relationship between conductivity data and impurity concentration is accurately fitted.

[0079] The category training samples are input into the classification module, and the support vector machine algorithm SVM separates the data of different categories by constructing a hyperplane. The hyperplane is expressed as:

[0080] w·x+b=0 (1)

[0081] Among them, w is the normal vector, b is the bias term, and x is the input data.

[0082] SVM uses the hinge loss function, which is defined as:

[0083]

[0084] in is the i-th category training sample in training set 1 The predicted value of Represents the category label of the i-th category training sample.

[0085] The goal of classification module training is to minimize the following loss function:

[0086]

[0087] Among them, C is the regularization parameter used to control the degree of fit between the complexity of the classification module and the training set, and n represents the number of class training samples in training set 1. In this way, SVM can learn the optimal classification boundaries for different impurity types.

[0088] The concentration fitting module uses a regression analysis algorithm to calculate the concentration of each impurity:

[0089]

[0090] Where Y is the concentration of each impurity, represents the j-th concentration training sample in training set 2, β0,β1,…,β m is the regression coefficient, and ε is the error term. In order to determine the optimal model parameters β0, β1,…, β m , the model is trained using the least squares method or gradient descent method.

[0091] The goal of concentration fitting module training is to minimize the loss function J(θ): J(θ) uses mean square error MSE as the loss function.

[0092]

[0093] Where m is the total number of concentration training samples in training set 2, is the actual impurity concentration of the j-th concentration training sample, is the predicted value of the concentration fitting module.

[0094] After multiple rounds of iterations, the model finally determined the optimal parameters β0, β1,…, β m, the model can accurately fit the relationship between conductivity and impurity concentration. The final regression analysis model can be applied to the online monitoring device to predict the concentration of each impurity in the transformer oil by inputting real-time conductivity data. The model can accurately reflect the relationship between conductivity and impurity concentration, ensuring the accuracy and reliability of online detection of particulate matter in transformer oil.

[0095] During the training process, the gradient descent algorithm is used to optimize the parameters of the classification module and the concentration fitting module, and to update the weight vector of the SVM and the coefficient of the regression analysis algorithm. The formula is as follows:

[0096]

[0097] Among them, θ is the model parameter, which is the weight vector in SVM and the regression coefficient in regression analysis algorithm; α is the learning rate, is the gradient of the loss function with respect to the parameters.

[0098] In the structure of the discriminant calculation model, the classification model and the concentration fitting model work closely together. The classification model determines the type of impurities in the oil sample based on the input sensor data and outputs a specific impurity category label as the basis for subsequent concentration calculations. After receiving the impurity type label, the concentration fitting model uses the spectrum and conductivity data to accurately calculate the concentration of the corresponding impurity. The organic combination of the two constitutes the core function of the discriminant calculation model, enabling it to comprehensively output the type of impurities and their concentrations.

[0099] After the training is completed, the entire discriminant calculation model is verified and optimized to ensure its accuracy and robustness, and the model is finally written into the microprocessor. In practical applications, the microprocessor relies on the discriminant calculation model to realize real-time monitoring and analysis of impurities in transformer oil, accurately determine the types of impurities in the oil and accurately calculate their concentration, thereby providing reliable data support for transformer status monitoring and maintenance.

[0100] Specifically, in the offline training stage, the method of using the multivariate linear regression algorithm and the principal component analysis algorithm to process the spectral data to calculate the concentration of different impurities is:

[0101] D concentration2 =X·β+ε1 (7)

[0102] Among them, D concentration1 is the impurity concentration, X is the spectral data obtained by the spectral sensor, β is the regression coefficient, and ε1 is the error term.

[0103] Specifically, in the offline training stage, the conductivity data is processed using a gradient descent algorithm to calculate the concentration of different impurities. The method is:

[0104] The concentration fitting model uses a linear regression algorithm to fit the relationship between conductivity data and impurity concentration:

[0105] D concentration2 =β0+β1·x1+β2·x2+…+β n ·x n (8)

[0106] Among them, D concentration2 is the impurity concentration 2, x1, x2,…, x n is the conductivity data obtained by the conductivity sensor, β0, β1,…, β n is the regression coefficient, and the regression coefficient is obtained by least squares method or gradient descent algorithm in the offline training stage.

[0107] Specifically, in the offline training stage, the concentration D calculated from the spectral sensor data of different impurities is concentration1 The concentration calculated from the conductivity sensor data is D concentration2 Perform fusion to obtain the final concentration of each impurity. The fusion formula is as follows:

[0108] D combined =α·D concentration1 +(1-α)·D concentration2 (9)

[0109] Among them, D combined It represents the final concentration of each impurity, α is the fusion weight, and the specific value can be adjusted according to the accuracy and characteristics of different sensors.

[0110] The present invention also provides a method for using a transformer oil particle online monitoring device based on spectrum and conductivity detection, comprising:

[0111] Step 1: Install and connect components

[0112] The staff first installed the sampling module (a micro pump model MP-100 and an adjustable flow valve model FV-200) at the sampling port of the transformer tank to ensure that the connection is stable and leak-free. Then, the sensor module (spectral sensor model SP-300 and conductivity sensor model EC-100) is connected to the output of the sampling module, and the analysis and processing unit (microprocessor MCU-500) and communication module (model: WT-400) are connected to ensure that the system connection is complete. Preferably, all connecting parts use oil-resistant materials to prevent leakage and contamination.

[0113] Step 2: Start the equipment and collect oil samples

[0114] Start the micro pump (MP-100), preferably, set the flow rate at 0.5L / min, and accurately control the flow rate through the adjustable flow valve (FV-200) to ensure the stability and continuity of the oil sample. The oil sample is pumped to the sensor module through the sampling module. Preferably, use a filter to filter the oil sample to avoid particles clogging the pipeline.

[0115] Step 3: Sensor module detection

[0116] Before starting the test, initialize the spectrum sensor (model: SP-300) and conductivity sensor (model: EC-100) to ensure that the sensors are working in the best condition. Preferably, use a known standard oil sample to calibrate the sensor and check the accuracy of the sensor response.

[0117] After the system is started, the sensor module continuously monitors the oil sample, and the analysis and processing unit (model: MCU-500) processes the data in real time to generate an accurate detection report. Preferably, in the offline stage, the spectral sensor performs optical detection on the oil sample within the wavelength range of 200-800nm, and uses a multivariate linear regression algorithm and a principal component analysis algorithm to process the spectral data to calculate the concentration of each different impurity, which is recorded as impurity concentration one; the conductivity sensor measures the conductivity change of the oil sample in the range of 0-200μS / cm, and uses a gradient descent algorithm to process the conductivity data to calculate the concentration of each different impurity, which is recorded as impurity concentration two. The impurity concentration one and the impurity concentration two corresponding to each impurity are fused to obtain the final concentration of each impurity; several final concentrations of each impurity are used as training set one, and the type of impurity in training set one is used as the label of the type training sample. The trained discriminant calculation model is written into the microprocessor MCU-500.

[0118] Further, the spectral sensor (model: SP-300) performs optical detection on the oil sample in the wavelength range of 200-800nm ​​to identify the spectral features in the oil sample. Preferably, the initial setting of the spectral sensor is performed using a calibration standard. A conductivity sensor (model: EC-100) is used to monitor the conductivity change of the oil sample, with a measurement range of 0-200μS / cm, reflecting the content of conductive impurities in the oil sample and providing supplementary electrical detection data. The conductivity sensor should be calibrated regularly to maintain accuracy.

[0119] Step 4: Data processing and analysis

[0120] In the online stage, the data information collected by the spectral sensor (model: SP-300) and the conductivity sensor (model: EC-100) is transmitted to the analysis and processing unit (model: MCU-500) in real time. The discriminant calculation model trained in the offline stage is used for processing, and the impurity type and concentration in the oil sample are judged.

[0121] Furthermore, the analysis and processing unit (model: MCU-500) integrates all analysis results, generates a test report, and stores the results in a built-in storage unit. Preferably, redundant storage technology is used to ensure data security and reliability.

[0122] Step 5: Data transmission

[0123] The communication module (model: WT-400) sends the processed data to the monitoring center in real time through wireless transmission technology (preferably using Wi-Fi or LoRa communication). The monitoring center checks the real-time detection results through the data monitoring platform to understand the operating status of the transformer. Preferably, the communication module should be configured with multiple encryption algorithms to ensure the security of data transmission.

[0124] Step 6: System maintenance and calibration

[0125] The system regularly performs self-check procedures to check the operating status of each component to ensure trouble-free operation. Preferably, the system is configured with an automatic alarm function, which automatically notifies maintenance personnel to check and handle when an abnormality is detected. Automatic self-check includes detecting the connection status of the component, the power status, the working status of the sensor, etc.

[0126] Preferably, the following operations are performed regularly to ensure long-term stable operation of the system:

[0127] Calibrate the spectroscopic sensor (model: SP-300) and the conductivity sensor (model: EC-100) to ensure detection accuracy. Preferably, a standard oil sample with known characteristics is used for calibration.

[0128] Check the working condition of the micro pump (MP-100) and the adjustable flow valve (FV-200), and perform maintenance and replacement when necessary to ensure the continuity and stability of oil sample collection.

[0129] Update the microprocessor algorithm to ensure the accuracy and efficiency of data processing. Preferably, update the software regularly to optimize the algorithm performance.

Claims

1. A transformer oil particle online monitoring device based on spectroscopy and conductivity detection, characterized in that: Includes the following components: Sampling module: including a micro pump and an adjustable flow valve, wherein the micro pump is used to realize continuous oil sample collection, and the adjustable flow valve is used to ensure the stability of the oil sample; Sensor module: including a spectral sensor and a conductivity sensor, wherein the spectral sensor is used to detect the optical properties in the oil sample, and the conductivity sensor is used to monitor the conductivity change of the oil sample and detect the electrical properties of different impurities in the oil; Analysis and processing unit: integrated microprocessor method, the microprocessor is used to process sensor data and determine the type and concentration of impurities; Communication module: wireless transmission technology is used to send the impurity type and concentration obtained by the analysis and processing unit to the monitoring center in real time.

2. The on-line monitoring device for particles in transformer oil according to claim 1, characterized in that: The integrated microprocessor includes an impurity type concentration discrimination method, the impurity type concentration discrimination method includes a discrimination calculation model, the discrimination calculation model includes a classification module and a concentration fitting module, wherein the classification module is used to determine the impurity type in the oil sample, and the concentration fitting module is used to calculate the impurity concentration corresponding to the impurity under the determined impurity type; The discrimination calculation model includes an offline training phase and an online detection phase; Offline training stage: Collect a number of offline data, including spectral data obtained by detecting different impurities in transformer oil through a spectral sensor and conductivity data obtained by detecting transformer oil under different impurity concentration conditions through a conductivity sensor, and label the spectral data and conductivity data of each impurity; The spectral data are processed by using a multivariate linear regression algorithm and a principal component analysis algorithm to calculate the concentration of each different impurity, which is recorded as impurity concentration one; the conductivity data are processed by using a gradient descent algorithm to calculate the concentration of each different impurity, which is recorded as impurity concentration two; then, the impurity concentration one and the impurity concentration two corresponding to each impurity are merged to obtain the final concentration of each impurity; several final concentrations of each impurity are used as training set one, and the types of impurities in training set one are used as labels of type training samples; A number of conductivity data of each impurity are used as training set 2, and the actual concentration of each impurity in training set 2 is used as the label of the corresponding concentration training sample; The classification module is trained through training set one, and when the loss no longer decreases, a trained classification module is obtained. The concentration fitting module is trained through training set two, and when the loss no longer decreases, a trained concentration fitting module is obtained. After the classification module and concentration fitting training are completed, a trained discrimination calculation model is obtained. Online detection stage: The data information collected by the spectral sensor and the conductivity sensor is input into the trained discrimination calculation model to output the types and concentrations of different particles.

3. The on-line monitoring device for particles in transformer oil according to claim 2, characterized in that: In the offline training stage, the classification module uses the support vector machine (SVM) algorithm to construct a hyperplane to separate different categories of impurity data and minimize the hinge loss function; the concentration fitting module uses a regression analysis algorithm to optimize the model parameters through the least squares method or gradient descent method to ensure that the relationship between conductivity data and impurity concentration is accurately fitted; The category training samples are input into the classification module, and the support vector machine algorithm SVM separates the data of different categories by constructing a hyperplane. The hyperplane is expressed as: w·x+b=0 (1) Among them, w is the normal vector, b is the bias term, and x is the input data; SVM uses the hinge loss function, which is defined as: in is the i-th category training sample in training set 1 The predicted value of Represents the category label of the i-th category training sample; The goal of classification module training is to minimize the following loss function: Among them, C is the regularization parameter, which is used to control the degree of fit between the complexity of the classification module and the training set, and n represents the number of category training samples in the training set 1; The concentration fitting module uses a regression analysis algorithm to calculate the concentration of each impurity: Where Y is the concentration of each impurity, represents the j-th concentration training sample in training set 2, β0,β1,…,β m is the regression coefficient, ε is the error term; The goal of concentration fitting module training is to minimize the loss function J(θ): Where m is the total number of concentration training samples in training set 2, is the actual impurity concentration of the j-th concentration training sample, is the predicted value of the concentration fitting module; During the training process, the gradient descent algorithm is used to optimize the parameters of the classification module and the concentration fitting module, and to update the weight vector of the SVM and the coefficient of the regression analysis algorithm. The formula is as follows: Among them, θ is the model parameter, which is the weight vector in SVM and the regression coefficient in regression analysis algorithm; α is the learning rate, is the gradient of the loss function with respect to the parameters.

4. The on-line monitoring device for particles in transformer oil according to claim 3, characterized in that: In the offline training stage, the method of using the multivariate linear regression algorithm and the principal component analysis algorithm to process the spectral data to calculate the concentration of different impurities is as follows: D concentration2 =X·β+ε1 (7) Among them, D concentration1 is the impurity concentration, X is the spectral data obtained by the spectral sensor, β is the regression coefficient, and ε1 is the error term.

5. The on-line monitoring device for particles in transformer oil according to claim 4, characterized in that: In the offline training stage, the conductivity data is processed using a gradient descent algorithm to calculate the concentration of different impurities. The method is: The concentration fitting model uses a linear regression algorithm to fit the relationship between conductivity data and impurity concentration: D concentration2 =β0+β1·x1+β2·x2+…+β n ·x n (8) Among them, D concentration2 is the impurity concentration 2, x1, x2,…, x n is the conductivity data obtained by the conductivity sensor, β0, β1,…, β n is the regression coefficient.

6. The on-line monitoring device for particles in transformer oil according to claim 5, characterized in that: In the offline training phase, the concentrations of different impurities calculated by the spectral sensor data are calculated. concentration1 The concentration calculated from the conductivity sensor data is D concentration2 The final concentration of each impurity is obtained by fusion. The fusion formula is as follows: D combined =α·D concentration1 +(1-a)·D concentration2 (9) Among them, D combined represents the final concentration of each impurity, and α is the fusion weight.

Images